Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 8 Jul 1997 18:24:52 -0500
From:      Raul Zighelboim <mango@staff.communique.net>
To:        "'hackers@freefall.freebsd.org'" <hackers@freefall.freebsd.org>
Subject:   There must be something really wrong with 2.2.2-release...
Message-ID:  <A03CD00C69B1D01195AB00A024ECEB161E317C@kaori.Communique.Net>

next in thread | raw e-mail | index | archive | help


Hello there.

	I manage a few FreeBSD boxes.
	3 of these boxes are running FreeBSD 2.2.2 RELEASE.Those 3 boxes
are configured very differently:

	1 Usenet news server with (a ppro with 256 megs of ram, used and
abused).
	1 a mail server with 18 megs of ram, and some 16 mailboxes
	1 a socks server with 64 megs of ram and a handfull (may 15 ! 20
at a time) connections.

The hardware on these 3 system is very different. they do not have
anything in common; different MB, controllers, ...

	These 3 boxes tend to crash after a while, with 'not enough swap
space' or not enough mbuf'.

Although I cannot guaranty that the problem is related to 2.2.2-RELEASE,
none of the other systems I have access to is having thes problems.

	Has anyine seens this behavior ?

As I monitor the news server much closely than the other machines:
	I have at this time 768Megs of swap space on the machine, and I
have noticed that:
	the first day the system is up, no more than 1% if the swap
space is used. The second day, I can see swap space jump to 60 - 70 -
80% (before I reboot the machine).  Killing all processes (no reboot)
extends the life of the system a little.
	I am running innd (with a few optimization patches) and named
uucp and sendmail on this system).



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?A03CD00C69B1D01195AB00A024ECEB161E317C>